ID 原文 译文
7274 相比于传统的旋转干涉仪测向方法,该方法可实现多目标的二维角度估计并且降低了对信噪比的要求。 Compared to the traditional rotary interferometer direction finding method, this method can realize multi-objective two-dimensional Angle estimation and reduces the requirement to the signal-to-noise ratio.
7275 针对基于卷积网络的超分辨率重建算法对不同场景下的图像存在复原质量不佳、细节信息丢失的问题,对卷积网络结构详细分析,结合重建模块和损失函数约束条件存在的问题,提出了基于并行映射卷积网络的超分辨率重建模型。 For super resolution reconstruction algorithm based on convolution network for different scenarios of poor image reconstruction quality and detail information leakage problems, detailed analysis of convolution network structure, combining with reconstruction module and loss function constraint problems, puts forward the model of super-resolution reconstruction based on parallel convolution mapping network.
7276 该模型基于端到端的思想,构建并行映射网络及正则化约束条件,能对图像特征进行层次化自主提取,在高分辨率图像重建时极大地丰富图像特征的维数; The model based on end-to-end, build parallel mapping network and regularization constraint conditions, can be hierarchical independent of image features are extracted, the high resolution image reconstruction greatly enrich the dimensions of the image characteristics;
7277 并且将全变分正则化引入到重建模块,有效地克服了超分辨率的病态问题,从而获得鲁棒、丰富的图像信息,提升了重建图像的质量。 And will be introduced to the total variation regularization reconstruction module, effectively overcome the super-resolution of ill-posed problems, to obtain robust, rich image information, improve the quality of the reconstructed image.
7278 实验结果表明,所提出的网络模型具有更优异的性能,其超分辨率算法在视觉评价和量化指标上取得了更好的重建效果。 The experimental results show that the proposed network model has more excellent performance, the super-resolution algorithm on the visual evaluation and quantitative indicators better reconstruction effect is obtained.
7279 针对磁屏蔽性能理论计算存在的不足,提出了基于径向基函数(radial basis function,RBF)神经网络的磁屏蔽性能理论计算方法。 For the theoretical calculation of magnetic shielding performance is proposed based on radial basis function (radial basis function, RBF) method to calculate the magnetic shielding performance of the neural network theory.
7280 首先,采用控制变量法对影响磁屏蔽性能的独立参数进行分离并建模; First of all, using the method of control variables affecting the magnetic shielding performance of independent parameters for separation and modeling;
7281 然后,利用RBF神经网络对非独立参数进行建模; Then, using RBF neural network for the independent parameter modeling;
7282 最后,将训练好的RBF神经网络模块与分离出来的独立参数模型进行结合,得到磁屏蔽装置的磁屏蔽性能计算模型。 Finally, the trained RBF neural network module combined with isolated independent parameters of the model, get the magnetic shielding device of magnetic shielding performance calculation model.
7283 通过对矩形磁屏蔽装置的磁屏蔽性能进行仿真计算。 Based on rectangular magnetic shielding performance of the magnetic shielding device simulation.